BTRFS major fault

BTRFS major fault

[regomodo]

I was running linux-kernel-2.6.32.1 and suspended my system using
"hibernate-ram".

When I resumed the system was OK for a moment then froze up. I managed to
get to tty1 easily and saw errors relating to btrfs.
I did not manage to get the error down but it had an abbreviation of
checksum and said maybe something about an inode/node. There was maybe 2
numbers possibly separated by  ":". I'm pretty sure the numbers were the
same at each stderr printout. Sorry for the vagueness.

Anyway. I tried to "reboot" but the errors continued and the shutdown
process wasn't happening. I r.e.i.s.u.b and reboot back into my system

The btrfs FS/volumes wouldn't mount so I tried to run a btrfsck. This is
what I got:

funtoo-pc ~ # btrfsck /dev/sda
failed to read /dev/sdf
failed to read /dev/sr0
btrfsck: disk-io.c:739: open_ctree_fd: Assertion `!(!tree_root->node)'
failed.
Aborted
funtoo-pc ~ # btrfsck /dev/sdb
failed to read /dev/sdf
failed to read /dev/sr0
btrfsck: disk-io.c:739: open_ctree_fd: Assertion `!(!tree_root->node)'
failed.
Aborted

ENV: funtoo-amd64; linux-2.6.32.1; x1 Btrfs FS, made up of x2 Samsung 500GB
sata hdd, in Raid0, and contained x2 subvolumes

I'm not trying to fix my issue -- I have a backup -- just reporting
something I found.


Cheers

Reg

Re: BTRFS major fault

[Chris Mason]

On Wed, Dec 16, 2009 at 06:29:38PM +0000, regomodo wrote:
> I was running linux-kernel-2.6.32.1 and suspended my system using
> "hibernate-ram".
> 
> When I resumed the system was OK for a moment then froze up. I managed to
> get to tty1 easily and saw errors relating to btrfs.
> I did not manage to get the error down but it had an abbreviation of
> checksum and said maybe something about an inode/node.

These show disk corruption, which is also what the btrfsck errors
report.  How exactly are you doing suspend?

-chris

Re: BTRFS major fault

[regomodo]

On Wednesday 16 December 2009 19:37:48 Chris Mason wrote:
> On Wed, Dec 16, 2009 at 06:29:38PM +0000, regomodo wrote:
> > I was running linux-kernel-2.6.32.1 and suspended my system using
> > "hibernate-ram".
> >
> > When I resumed the system was OK for a moment then froze up. I managed to
> > get to tty1 easily and saw errors relating to btrfs.
> > I did not manage to get the error down but it had an abbreviation of
> > checksum and said maybe something about an inode/node.
> 
> These show disk corruption, which is also what the btrfsck errors
> report.  How exactly are you doing suspend?
> 
>From the command-line, as root, with "hibernate-ram".